Fix: Attach _linear_extra_repr to quantized linear layers #1546
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/torchchat/1546
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
Thank you for your pull request and welcome to our community. Action RequiredIn order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you. ProcessIn order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA. Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with If you have received this in error or have any questions, please contact us at [email protected]. Thanks! |
I've made changes to resolve the AttributeError: 'Linear' object has no attribute '_linear_extra_repr' during quantization by adding necessary imports—import types and from torchao.quantization.quant_api import _linear_extra_repr—to torchchat/utils/quantize.py, enabling dynamic method attachment and access to _linear_extra_repr. Additionally, I modified the quantize_model function by introducing a recursive helper function _attach_extra_repr within the same file. This function traverses all modules in the model and, for each nn.Linear module missing the extra_repr attribute, dynamically attaches _linear_extra_repr using types.MethodType. The helper is invoked at the beginning of quantize_model to ensure all nn.Linear modules have the extra_repr method before quantization begins.